The broad objective is to provide a more detailed depiction of the cognitive and neural systems involved in verbal short-term (working) memory. A widely accepted view of verbal working memory is that it consists of a short-term store for phonological codes and an articulatory rehearsal process that refreshes the contents of the store (the phonological loop). Neither sensory-specific codes nor lexical-semantic codes are thought to be actively involved in the retention process. Lexical-semantic codes are thought to contribute to verbal working memory at the time of retrieval. An alternative view is that retention of verbal material in working memory involves the parallel operation of the phonological loop and memory traces based upon: 1) automatically maintained modality-specific sensory codes that retain a partial record of recent sounds and sights, and 2) lexical-semantic codes in long-term memory that are activated by attention focused upon them by the retention process, and in turn are the basis for an episodic record of the material held in working memory. It is difficult to decide between these different conceptions of memory from behavioral data alone, since behavioral data reflect a combination of retention and retrieval operations. The approach will be to utilize the high temporal resolution of event-related potentials (ERPs) recorded from human scalp to index brain activity specific to retention. Two experiments will deal with how sensory codes enter into retention of verbal material. A third experiment seeks information on semantic processing during retention of related and unrelated words. These experiments will rely upon ERP slow waves to depict spatio-temporal patterns of brain activation. Two more experiments will rely upon an ERP component that is sensitive to priming (N400) to assess whether retention of words increases activation of their semantic and phonological codes in long-term memory. Working memory deficits are a key aspect of many cognitive-neurological disorders. The efficacy of rehabilitative therapies for cognitive disorders depends upon a detailed functional analysis. The understanding of working memory provided by ERPs can contribute to identification of impaired components of the cognitive system, which is essential for developing and monitoring therapeutic treatments. As knowledge of the relationship between ERPs and working memory increases, this methodology could become a vital noninvasive tool for diagnosis and therapy.